EGU24-18285, updated on 11 Mar 2024
https://doi.org/10.5194/egusphere-egu24-18285
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Large-scale remote monitoring of riverine litter to evaluate effectiveness of clean-up technologies

Liesbeth De Keukelaere, Els Knaeps, Robrecht Moelans, Marian-Daniel Iordache, Klaas Pauly, and Ils Reusen
Liesbeth De Keukelaere et al.
  • VITO, Environmental Intelligence, Mol, Belgium (liesbeth.dekeukelaere@vito.be)

In June 2023 the Horizon Europe project INSPIRE kicked off. INSPIRE will fight against the plastic pollution in rivers by introducing 20 scalable technologies to prevent and eliminate litter. The technologies will be demonstrated in 6 rivers across Europe. Monitoring of the plastic load in the river and the riverbanks is essential to develop a baseline and evaluate effectiveness of the technologies. Here we will introduce a camera and drone-based system to monitor plastic flux in the river and macroplastic densities on the riverbanks. The fixed camera system consists of a series of Commercial Of-The-Shelf (COTS) surveillance cameras with housing and real-time datalink. The cameras work autonomous and will provide a continuous feed of data uploaded to the cloud. The drone system consists of a high resolution RGB and multispectral Micasense camera. Specific attention goes to the conversion from the raw drone data into standardized Analysis Ready Data (ARD) including: (1) image alignment of the multispectral camera. (2) Converting raw drone data into reflectance products (using an irradiance sensor) allows the methodology to be applicable in any circumstance (clear, overcast, cloudy conditions) and transferable to other regions. (3) Sensor fusion, to align high spatial resolution RGB with high spectral resolution MicaSense data.

 

For plastic detection and characterization robust machine learning models are being used including new pre-trained foundation models like Segment Anything. New methods are being tested to transform the camera-based plastic detections into a plastic flux product taking into account the tide effects in the river. This includes feature detection techniques like SIFT (Scale_Invariant Feature Transform), SURF (Speeded-Up Robust Features) or ORB (Oriented FAST and Rotated Binary Robust Independent Elementary Features) in combination with a feature matching algorithm (e.g. FLANN based matcher). Here, we will present the INSPIRE project and its first results demonstrated at the Temse bridge (Belgium) and riverbanks along the Scheldt river (Belgium).

How to cite: De Keukelaere, L., Knaeps, E., Moelans, R., Iordache, M.-D., Pauly, K., and Reusen, I.: Large-scale remote monitoring of riverine litter to evaluate effectiveness of clean-up technologies, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18285, https://doi.org/10.5194/egusphere-egu24-18285, 2024.

Supplementary materials

Supplementary material file

Comments on the supplementary material

AC: Author Comment | CC: Community Comment | Report abuse

supplementary materials version 1 – uploaded on 16 Apr 2024, no comments

Post a comment